How to Pinpoint the Exact Java Code Causing 100% CPU Usage with jstack and Arthas

Why CPU 100% Happens and How to Diagnose

CPU spikes in production often stem from Java code that consumes all CPU. Being able to locate the exact line of code is a key skill for interviews and troubleshooting.

Two Main Tools

jstack – the JVM‑provided thread‑stack dump utility.

Arthas – an open‑source Java diagnostic tool that also provides thread and stack information.

Using jstack

First find the process ID (PID) of the Java program, then run top -Hp <pid> to see which thread uses the most CPU. Convert the thread’s decimal ID to hexadecimal with printf "%x\n" <tid>, then dump the stack with jstack -l <pid>. The output shows each thread, its state, and the source line (e.g., JVMCPU.java:16) that is consuming CPU.

Example command sequence:

top -Hp 32208
printf "%x
" 32230   # => 0x7de6
jstack -l 32208

In the stack dump, look for the line containing JVMCPU$CPUTask.run(JVMCPU.java:16). This points to the busy loop in the sample code.

Sample Java Program

package com.tuling.learnjuc.demo;

import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;

public class JVMCPU {
    private static ExecutorService service = Executors.newFixedThreadPool(5);
    private static Object lock = new Object();

    public static class CPUTask implements Runnable {
        @Override
        public void run() {
            synchronized (lock) {
                long sum = 0L;
                while (true) {
                    sum += 1;
                }
            }
        }
    }

    public static void main(String[] args) {
        CPUTask cpuTask = new CPUTask();
        service.execute(cpuTask);
    }
}

Using Arthas

Download and start Arthas:

curl -O https://arthas.aliyun.com/arthas-boot.jar
java -jar arthas-boot.jar

In the Arthas console run dashboard to list top‑N threads, then thread -n 1 to print the stack of the busiest thread. The result is similar to jstack and also points to JVMCPU.java:16.

Both tools ultimately reveal that the infinite loop in while(true){ sum+=1; } is the cause of the CPU 100% usage.